Project 2: Predicting surgical skills from fNIRS via deep learning models

I achieved an accuracy of R2 = 0.73 and AUC = 0.91 by designing a convolutional neural network (CNN) model to extract features from fNIRS data to regress out the motor skill level, which is much higher than conventional machine learning models, including support vector regression (SVR), kernel partial least squares (KPLS) and random forest (RF) algorithms.